A functional analysis of PCNA-binding peptides derived from protein sequence, interaction screening and rational design

Abstract

Proliferating cell nuclear antigen (PCNA) has no intrinsic enzymatic function, but functions as a sliding platform to mediate protein interactions with the DNA strand. Many proteins interact with PCNA through a small conserved motif with consensus QxxLxxFF. This work uses Schizosaccharomyces pombe and human cells to analyse the function of PCNA-binding peptides. Interacting peptides were identified using two-hybrid screening; one (pep102) binds directly to a physiologically relevant site on PCNA. The EGFP-pep102 overexpression phenotype is consistent with competitive blocking of PCNA-protein interactions. Various PCNA-binding peptides were all shown to inhibit PCNA function by competitive binding in both human and S. pombe cells as EGFP fusion proteins. The action of a p21(WAF1/Cip1)-derived peptide was complicated by the presence of additional functional domains and possible post-translational modification. The activity of pep102 was hampered by low expression in both model systems. The peptide derived from rational design (con1) was stable, highly active in inhibiting PCNA function both S. pombe and human cells and showed a high affinity for PCNA both in vitro and in vivo. These results validate the use of functional screening in yeast to identify peptide aptamers that are functional in mammalian cells; such aptamers provide excellent leads for small molecule antiproliferative therapies.

Figures

Figure 1. In vitro and in vivo analysis of peptide-PCNA interactions

Panel A: Results of β-galactosidase assays on two-hybrid transformants. p102h and p103h are as described in the text; pACT-PCNA expresses full length S. pombe PCNA as a fusion with Gal4ACT pACT-SNF expresses yeast SNF4 as a control and pGAD-GH is the library vector. These were tested against PCNA from various species expressed as fusions with Gal4AS. “Lamin” was used as a negative control. Panel B: ELISA assays of human PCNA binding to peptides. Peptides were immobilised on streptavidin-coated multiwell plates through an N-linked biotin moiety (as indicated) and binding of purified human PCNA detected using two anti-PCNA antibodies, 3009 or PC10. Control indicates an unrelated peptide and none indicates that no peptide was used for affinity capture. Panel C: Affinity capture experiments were carried out from HeLa cell extracts with peptides as shown; bound PCNA was detected by Western blot as indicated by the arrow.

Figure 2. Peptides share a common binding site on human PCNA

The ability of bead-conjugated peptides to bind to PCNA in HeLa cell extracts was tested in the presence of varying concentrations of free, unbiotinylated peptides as described in Materials and Methods. The relative amounts of bound PCNA are shown by Western blot analysis. In panels A, B and C the ability of PCNA in HeLa cell extracts to bind to 102 conjugated beads was tested, in the presence of various competing peptides. In panels D, E and F, the ability of PCNA to bind to pogo-conjugated peptide beads was tested. The addition of an unrelated peptide (control) did not affect the biotinylated peptide-PCNA interaction.

Figure 3. Western blot analysis of EGFP-peptide constructs expressed in vivo

Peptide-EGFP fusion proteins were expressed in human U2OS cells and cell lysates prepared as described. Immunoprecipitation was carried out with the polyclonal anti-EGFP antibody Ab290 (AbCam) and the precipitated proteins analysed by SDS-PAGE and Western blot. “EGFP” indicates that EGFP was expressed from the vector with no additional fusion sequence; “none” indicates transfection with an unrelated plasmid. Other EGFP fusions with peptides are as shown. The amounts of plasmid DNA used for transfection of a 10cm dish were as follows: pEGFP-con1: 10μg; pEGFP-pep10: 10μg; pEGFP-102: 30μg; pEGFP-mut10: 2.5μg; pEGFP: 10μg. In the upper panel, input levels of pEGFP fusion proteins are shown in a Western blot probed with a mouse monoclonal anti-EGFP antibody (Roche). The middle panel shows input levels of PCNA. The lower panel shows Western blot analysis of PCNA co-immuoprecipitated with pEGFP fusion proteins. The lower band in this panel is cross-reactivity of the antibody light chain.

Figure 4. The effects of EGFP-pep102 expression on cell number and viability in S. pombe

Protein expression was induced in S. pombe cells containing integrated constructs expressing either EGFP or EGFP-pep102 by inoculating log phase cells into thiamine-free medium. Cell numbers (A) and viability (B) were determined and the results of a representative experiment shown. “rad1” indicates the transformants had a rad1-1 genetic background.

Figure 5. Expression of EGFP-pep102 in S. pombe results in a checkpoint-dependent cell cycle arrest

Following induction of protein expression, cells expressing EGFP or EGFP-pep102 were harvested, fixed and stained with DAPI as described. The results of expressing EGFP-pep102 are shown (right column) compared to those of EGFP alone (left column) either in a rad1+ (upper row) or a rad1-1 genetic background (lower row).

Figure 6. Levels of pEGFP-fusion proteins and PCNA expressed in S. pombe

Upper panel shows a comparison of endogenous PCNA with levels in dilutions of protein extracts from cells over-expressing PCNA. Protein expression was induced for 24 hours in cells containing either vector (pREP1) or pREP1-pcn1+, which expresses S. pombe PCNA under the control of the nmt promoter. Soluble protein extracts were prepared and analysed by SDS-PAGE and Western blot analysis using the anti-PCNA antibody PC10. “E” indicates endogenous PCNA present in the vector-only transformants. In the remaining five lanes, increasing amounts of protein extract from pREP1-pcn1+ transformants was loaded at dilutions of 1 in 500, 1 in 400, 1 in 300, 1 in 200 and 1 in 100. The lower panel shows levels of EGFP and EGFP-peptide fusion proteins expressed with (+) and without (−) co-expressed PCNA. Protein expression (as above) was induced for 24 hours. The panel showing EGFP-102 expression has been exposed approximately 8 times longer than the main panel.

Figure 7. The effects of EGFP-peptide expression upon cell number in S. pombe

Following induction of protein expression by inoculation into medium lacking thiamine, cell numbers were determined for transformants expressing various EGFP-peptide fusion proteins as shown. 4xG indicates EGFP expression with no fusion from the vector. The results of a representative experiment are shown.

Figure 8. EGFP-peptide fusions inhibit colony formation in human tumour cell lines

Clonogenic assays were carried out to assess the effects of EGFP-peptide fusions – for details see Materials and Methods. Transfected cells were grown for 10 to 20 days under selection before fixation and staining with Giemsa. For each pair of wells, the left hand wells were seeded with twice as many cells as the right. Representative results are shown in each case for HeLa (cervical cancer) and for U2OS (osteosarcoma) cell lines.